process_data.py

import pandas as pd
import talib
import numpy as np
from sklearn.preprocessing import MinMaxScaler
import ta
from ta import zigzag, money_flow_index

class FeatureExtractor:
    def __init__(self, df):
       self.df = df
       self.open = df['open'].astype('float')
       self.close = df['close'].astype('float')
       self.high = df['high'].astype('float')
       self.low = df['low'].astype('float')
       self.volume = df['volume'].astype('float')

    def add_bar_features(self):
        # stationary candle
        self.df['bar_hc'] = self.high - self.close
        self.df['bar_ho'] = self.high - self.open
        self.df['bar_hl'] = self.high - self.low
        self.df['bar_cl'] = self.close - self.low
        self.df['bar_ol'] = self.open - self.low
        self.df['bar_co'] = self.close - self.open
        self.df['bar_mov'] = self.df['close'] - self.df['close'].shift(1)
        return self.df


    def add_mv_avg_features(self):
        self.df['sma5'] = talib.SMA(self.close,5)
        self.df['sma20'] = talib.SMA(self.close,20)
        self.df['sma120'] = talib.SMA(self.close,120)
        self.df['ema12'] = talib.SMA(self.close,5)
        self.df['ema26'] = talib.SMA(self.close,26)
        return self.df

    def add_adj_features(self):
        self.df['adj_open'] = self.df['open'] / self.close
        self.df['adj_high'] = self.df['high'] / self.close
        self.df['adj_low'] = self.df['low'] / self.close
        self.df['adj_close'] = self.df['close'] / self.close
        return self.df

    # note! this is not a complete list
    # additional indicator can help in some scenario but usually acts as a noise
    def add_ta_features(self):
        obv = talib.OBV(self.close, self.volume)
        obv_mv_avg = talib.MA(obv, timeperiod=10)
        obv_mv_avg[np.isnan(obv_mv_avg)] = obv[np.isnan(obv_mv_avg)]
        difference = obv - obv_mv_avg

        self.df['obv'] = obv
        self.df['obv_signal'] = difference
        self.df['obv_cheat'] = np.gradient(difference)

        upper, middle, lower = talib.BBANDS(self.close, timeperiod=20, nbdevup=2, nbdevdn=2, matype=0)

        self.df['dn'] = lower
        self.df['mavg'] = middle
        self.df['up'] = upper
        self.df['pctB'] = (self.close - self.df.dn) / (self.df.up - self.df.dn)
        rsi14 = talib.RSI(self.close, 14)
        self.df['rsi14'] = rsi14

        macd, macdsignal, macdhist = talib.MACD(self.close, 12, 26, 9)
        self.df['macd'] = macd
        self.df['signal'] = macdsignal

        ## addtional info
        self.df['adx'] = talib.ADX(self.high, self.low, self.close, timeperiod=14)
        self.df['cci'] = talib.CCI(self.high, self.low, self.close, timeperiod=14)

        ## maximum profit
        self.df['plus_di'] = talib.PLUS_DI(self.high, self.low, self.close, timeperiod=14)

        ## lower_bound
        self.df['lower_bound'] = self.df['open'] - self.df['low'] + 1

        ## ATR
        self.df['atr'] = talib.ATR(self.high, self.low, self.close, timeperiod=14)

        ## STOCH momentum
        self.df = ta.stochastic_oscillator_k(self.df)
        self.df = ta.stochastic_oscillator_d(self.df, n=10)

        ## TRIX
        self.df['trix'] = talib.TRIX(self.close, timeperiod=5)
        self.df['trix_signal'] = ta.moving_average(self.df['trix'], n=3)
        self.df['trix_hist'] = self.df['trix'] - self.df['trix_signal']

        ## MFI

        self.df['mfi14'] = money_flow_index(self.df, 14)